DRSEL: A dual branch feature level ensemble learning framework based on multi sensor data fusion for fault diagnosis of rotating machines

Yuan Zhuang; Wei Qu; Junjie Ding; Minling Pan; Jiahua Su

doi:10.59400/sv3354

DRSEL: A dual branch feature level ensemble learning framework based on multi sensor data fusion for fault diagnosis of rotating machines

Yuan Zhuang College of Mechanical, Naval Architecture & Ocean Engineering, Beibu Gulf University, Qinzhou 535011, China Guangxi Key Laboratory of Precision Navigation Technology and Application, Guilin University of Electronic Technology, Guilin 541004, China Shenzhen Junrongyao Precision Hardware Products Co., Ltd., Shenzhen 518000, China
Wei Qu College of Electronic and Information Engineering, Beibu Gulf University, Qinzhou 535011, China
Junjie Ding Shenzhen Junrongyao Precision Hardware Products Co., Ltd., Shenzhen 518000, China School of Information Science, Guangdong University of Finance & Economics, Guangzhou 510320, China
Minling Pan College of Electronic and Information Engineering, Beibu Gulf University, Qinzhou 535011, China
Jiahua Su College of Mechanical, Naval Architecture & Ocean Engineering, Beibu Gulf University, Qinzhou 535011, China

Article ID: 3354

Keywords: multi-sensor data fusion; feature-level ensemble learning; fault diagnosis; train bogie gearbox; induction motor

Abstract

This paper proposes a feature-level ensemble learning framework for fault diagnosis of rotating machinery based on multi-sensor data fusion, aiming to address the inherent limitations of conventional single-model or single-sensor approaches in capturing complex and variable fault characteristics. Firstly, an improved ensemble learner is designed by integrating residual networks with Swin Transformer blocks, enabling the extraction of multi-scale, hierarchical, and complementary features from multi-sensor vibration signals. This design allows the model to capture local fine-grained patterns while simultaneously learning long-range dependencies, thus achieving a more comprehensive and discriminative feature representation. Then, the feature vectors produced by multiple base learners are concatenated to perform feature-level fusion, which effectively leverages the complementary information across heterogeneous sensors and significantly enhances diagnostic accuracy, robustness, and stability. Finally, the proposed framework is validated through two real-world industrial case studies involving a train bogie gearbox and an induction motor, covering diverse operating speeds, load conditions, and fault types. Experimental results demonstrate that the method achieves a fault diagnosis accuracy exceeding 96.88%, markedly outperforming traditional single-model approaches and conventional fusion strategies. Moreover, the framework exhibits strong generalization ability under variable working conditions. These findings highlight its practical applicability in industrial scenarios and underline its potential to support the development of intelligent and reliable predictive maintenance systems.

Published

2025-09-01

How to Cite

Zhuang, Y., Qu, W., Ding, J., Pan, M., & Su, J. (2025). DRSEL: A dual branch feature level ensemble learning framework based on multi sensor data fusion for fault diagnosis of rotating machines. Sound & Vibration, 59(5). https://doi.org/10.59400/sv3354

Download Citation

Issue

Vol. 59 No. 5 (2025)

Section

Article

This work is licensed under a Creative Commons Attribution 4.0 International License.

References

[1]He DQ, Wu JX, Jin ZZ, et al. AGFCN: A bearing fault diagnosis method for high-speed train bogie under complex working conditions. Reliability Engineering & System Safety. 2025; 258: 110907. doi: 10.1016/j.ress.2025.110907

[2]He DQ, Zhao JY, Jin ZZ, et al. Prediction of bearing remaining useful life based on a two-stage updated digital twin. Advanced Engineering Informatics. 2025; 65: 103123. doi: 10.1016/j.aei.2025.103123

[3]Tama BA, Vania M, Lee S, et al. Recent advances in the application of deep learning for fault diagnosis of rotating machinery using vibration signals. Artificial Intelligence Review. 2023; 56(5): 4667–4709. doi: 10.1007/s10462-022-10293-3

[4]Zhu ZQ, Lei YB, Qi GQ, et al. A review of the application of deep learning in intelligent fault diagnosis of rotating machinery. Measurement. 2023; 206: 112346. doi: 10.1016/j.measurement.2022.112346

[5]Qin YL, He DQ, Jin ZZ, et al. An improved deep learning algorithm for obstacle detection in complex rail transit environments. IEEE Sensors Journal. 2024; 24(3): 4011–4022. doi: 10.1109/jsen.2023.3340688

[6]Sun HM, He DQ, Zhong JC, et al. Preventive maintenance optimization for key components of subway train bogie with consideration of failure risk. Engineering Failure Analysis. 2023; 154: 107634. doi: 10.1016/j.engfailanal.2023.107634

[7]Yin XH, Mu ZQ, Cui QA, et al. Interpretable and spatio-distributed settlement related multimode process monitoring for Metro tunnels excavated by TBM. Advanced Engineering Informatics. 2025; 66: 103464. doi: 10.1016/j.aei.2025.103464

[8]Lao ZP, He DQ, Wei ZX, et al. Intelligent fault diagnosis for rail transit switch machine based on adaptive feature selection and improved LightGBM. Engineering Failure Analysis. 2023; 148: 107219. doi: 10.1016/j.engfailanal.2023.107219

[9]Qin GH, Zhang K, Lai XW, et al. An adaptive symmetric loss in dynamic wide-kernel ResNet for rotating machinery fault diagnosis under noisy labels. IEEE Transactions on Instrumentation and Measurement. 2024; 73: 3517512. doi: 10.1109/tim.2024.3375404

[10]Hou SX, Lian A, Chu YD. Bearing fault diagnosis method using the joint feature extraction of Transformer and ResNet. Measurement Science and Technology. 2023; 34(7): 075108. doi: 10.1088/1361-6501/acc885

[11]Xiao YM, Shao HD, Wang J, et al. Bayesian variational transformer: A generalizable model for rotating machinery fault diagnosis. Mechanical Systems and Signal Processing. 2024; 207: 110936. doi: 10.1016/j.ymssp.2023.110936

[12]Lv J, Xiao QY, Zhai XD, et al. A high-performance rolling bearing fault diagnosis method based on adaptive feature mode decomposition and Transformer. Applied Acoustics. 2024; 224: 110156. doi: 10.1016/j.apacoust.2024.110156

[13]Yan S, Shao HD, Wang J, et al. LiConvFormer: A lightweight fault diagnosis framework using separable multiscale convolution and broadcast self-attention. Expert Systems with Applications. 2024; 237: 121338. doi: 10.1016/j.eswa.2023.121338

[14]Kibrete F, Woldemichael DE, Gebremedhen HS. Multi-sensor data fusion in intelligent fault diagnosis of rotating machines: A comprehensive review. Measurement. 2024; 232: 114658. doi: 10.1016/j.measurement.2024.114658

[15]Lao ZP, He DQ, Jin ZZ, et al. Few-shot fault diagnosis of turnout switch machine based on semi-supervised weighted prototypical network. Knowledge-Based Systems. 2023; 274: 110634. doi: 10.1016/j.knosys.2023.110634

[16]Zhuang Y, He DQ, Jin ZZ, et al. SWR2 Net: A fault diagnosis framework for rotating machine under limited samples and noise interference. Nonlinear Dynamics. 2025; 113(17): 22823–22852. doi: 10.1007/s11071-025-11302-0

[17]Wang SQ, Feng ZG. Multi-sensor fusion rolling bearing intelligent fault diagnosis based on VMD and ultra-lightweight GoogLeNet in industrial environments. Digital Signal Processing. 2024; 145: 104306. doi: 10.1016/j.dsp.2023.104306

[18]He DQ, Xu Y, Jin ZZ, et al. A zero-shot model for diagnosing unknown composite faults in train bearings based on label feature vector generated fault features. Applied Acoustics. 2025; 232: 110563. doi: 10.1016/j.apacoust.2025.110563

[19]Zhang YC, Ding JL, Li YB, et al. Multi-modal data cross-domain fusion network for gearbox fault diagnosis under variable operating conditions. Engineering Applications of Artificial Intelligence. 2024; 133: 108236. doi: 10.1016/j.engappai.2024.108236

[20]Qiu Z, Fan SF, Liang HB, et al. Multimodal fusion fault diagnosis method under noise interference. Applied Acoustics. 2025; 228: 110301. doi: 10.1016/j.apacoust.2024.110301

[21]Xu ZZ, Chen X, Xu JT. Multi-modal multi-sensor feature fusion spiking neural network algorithm for early bearing weak fault diagnosis. Engineering Applications of Artificial Intelligence. 2025; 141: 109845. doi: 10.1016/j.engappai.2024.109845

[22]Mian Z, Deng XF, Dong XH, et al. A literature review of fault diagnosis based on ensemble learning. Engineering Applications of Artificial Intelligence. 2024; 127: 107357. doi: 10.1016/j.engappai.2023.107357

[23]He YL, He DQ, Lao ZP, et al. Few-shot fault diagnosis of turnout switch machine based on flexible semi-supervised meta-learning network. Knowledge-Based Systems. 2024; 294: 111746. doi: 10.1016/j.knosys.2024.111746

[24]Tong JY, Liu C, Bao JH, et al. A novel ensemble learning-based multisensor information fusion method for rolling bearing fault diagnosis. IEEE Transactions on Instrumentation and Measurement. 2023; 72: 9501712. doi: 10.1109/tim.2022.3225910

[25]Ye MY, Yan XA, Jiang D, et al. MIFDELN: A multi-sensor information fusion deep ensemble learning network for diagnosing bearing faults in noisy scenarios. Knowledge-Based Systems. 2024; 284: 111294. doi: 10.1016/j.knosys.2023.111294

[26]Fu GH, Wang XG, Liu YH, et al. A robust bearing fault diagnosis method based on ensemble learning with adaptive weight selection. Expert Systems with Applications. 2025; 269: 126420. doi: 10.1016/j.eswa.2025.126420

[27]You YW, Tang JH, Guo M, et al. Ensemble learning based multi-fault diagnosis of air conditioning system. Energy and Buildings. 2024; 319: 114548. doi: 10.1016/j.enbuild.2024.114548

[28]Xiao YM, Shao HD, Wang J, et al. Domain-augmented meta ensemble learning for mechanical fault diagnosis from heterogeneous source domains to unseen target domains. Expert Systems with Applications. 2025; 259: 125345. doi: 10.1016/j.eswa.2024.125345

[29]He KM, Zhang XY, Ren SQ, et al. Deep Residual Learning for Image Recognition. In: Proceedings of the 2016 IEEE Conference on Computer Vision and Pattern Recognition; 27–30 June 2016; Las Vegas, NV, USA. pp. 770–778. doi: 10.1109/cvpr.2016.90

[30]Liu Z, Lin YT, Cao Y, et al. Swin Transformer: Hierarchical Vision Transformer using Shifted Windows. In: Proceedings of the 18th IEEE/CVF International Conference on Computer Vision; 10–17 October 2021; Montreal, QC, Canada. pp. 9992–10002. doi: 10.1109/iccv48922.2021.00986

[31]Ding A, Qin Y, Wang B, et al. Evolvable graph neural network for system-level incremental fault diagnosis of train transmission systems. Mechanical Systems and Signal Processing. 2024; 210: 111175. doi: 10.1016/j.ymssp.2024.111175

[32]Sehri M, Dumond P. University of Ottawa constant and variable speed electric motor vibration and acoustic fault signature dataset. Data in Brief. 2024; 53: 110144. doi: 10.1016/j.dib.2024.110144

This paper introduces a universal framework for understanding the vibration responses of systems subjected to harmonic excitation. By examining a simplified cylinder-spring-damper model, the study refurbishes traditional scaling methods for the excitation frequency ratio and critical damping ratio. The findings indicate that in damped systems, the maximum amplitude of vibration does not align with the natural frequency. This observation leads to the introduction of a new scaling method for reduced frequency. This new approach aligns resonance peaks at the new reduced velocity of 1.0 across different damping ratios, providing a consistent characterization of vibration behavior. A new critical damping ratio of 0.707 is identified for an excited system as opposed to the traditional damping ratio of 1.0 for an unexcited system. Key properties such as maximum amplitude, phase lag, bandwidth, and quality factor are analyzed, demonstrating that the proposed reduced frequency and critical damping ratio effectively capture the dynamics of both damped and undamped excited systems. The findings offer significant insights for practical applications in engineering and various scientific fields.

This study makes significant contributions to the field of ultrasonic testing (UT) by offering a novel approach to the identification of artificially introduced defects within Japanese cedar wood (Cryptomeria japonica). The findings are of particular relevance for the heritage conservation and construction sectors, where non-invasive defect detection is paramount. The study establishes a robust framework for assessing the structural integrity of timber by correlating ultrasonic wave velocity reductions with defect size and distribution. Big-sized defects led to more substantial decreases in wave velocity. The study establishes a robust framework for assessing the structural integrity of historical timber by correlating ultrasonic wave velocity reductions with defect size and distribution. This framework has the potential to be applicable to diverse wood species and defect types.

The control of vehicle interior noise has become a critical metric for assessing noise, vibration, and harshness (NVH) in vehicles. During the initial phases of vehicle development, accurately predicting the impact of road noise on interior noise is essential for reducing noise levels and expediting the product development cycle. In recent years, data-driven methods based on machine learning have gained significant attention due to their robust capability in navigating complex data mapping relationships. Notably, surrogate models have demonstrated exceptional performance in this domain. Numerous researchers have integrated diverse intelligent algorithms into the study of vehicle noise, leveraging advantages such as the elimination of precise modeling requirements, extensive solution space exploration, continuous learning from data, and robust algorithmic versatility. However, in NVH engineering applications, data-driven models face inherent limitations, particularly in interpretability and stability. To address these issues, this paper introduces an improved Long Short-Term Memory (LSTM) network that combines knowledge and data. Inspired by the physical information neural network concept, this approach incorporates values calculated through empirical formulas into the neural network as constraints. Comparative assessments with traditional LSTM networks highlight the advantages of this deep learning model. By integrating empirical formulas constraints, the model not only enhances interpretability but also achieves robust generalization with fewer data samples. The proposed method is validated on a specific vehicle model, showing significant improvements in prediction accuracy and efficiency.